GTM engineering is a new title for an old job

GTM engineering is the hottest job title in B2B right now, and the job itself is older than the CRM: remove the human bottlenecks between a buying signal and a sent email.

I run a GTM engineering operation. It sent 5 million+ cold emails in the past 12 months and booked 3,000+ calls for clients in the last year. Most of the guides ranking for this term are written by the tool vendors selling the stack, so consider this the view from inside a machine that actually runs.

What GTM engineering actually is

Strip the branding and GTM engineering is this: building the systems that spot a buyer showing intent, turn that signal into a precise list, and get a relevant message out the door without a human copying and pasting in the middle.

The title is new. Clay’s guide to GTM engineering tracks the role emerging around 2022, with job listings growing 205% between 2024 and 2025. Every VC content team has published an explainer. LinkedIn has anointed it the future of sales.

The job is not new. In 2016 the same person was called sales ops with a scraper habit. In 2020 they were called growth. The work has always been the same: signal comes in, outreach goes out, and every manual step between those two points is latency someone should remove.

What changed is the cost of building. One operator with Claude Code and a stack of API keys now ships internal tooling that took a four-person ops team in 2021. That collapse is the whole reason the title exploded. It is also the reason the role attracts so much theater, which we will get to.

What does a GTM engineer do all day

Most guides answer this with a responsibilities list written by someone who never held the pager. Here is what the role looks like inside our operation. The work comes down to 4 jobs:

1. Lead list construction.Not exporting a category from a database. A category is a bucket, and you can’t write to a bucket. The job is building lists around the specific human the filter can’t see, stacking title variations so the search doesn’t miss half the market (a founder is also a CEO, an owner, a managing partner, a president), and doing it cheaply. A 15,000-person lead list costs about $29 when you skip the enterprise data tools.
2. Signal monitoring. Watching for the events that mean a company is in motion: hiring posts, ad spend, social activity, funding. The list rebuilds itself around who is showing intent this week instead of who existed in a database last year.
3. Analytics. Pulling campaign data and answering operator questions: which buyer profile replies, which offer angle pulls positive responses, where volume is bleeding. We run this through Claude Code workflows we wrote ourselves, the same workflows that produce first-draft copy a human then edits.
4. Post-reply and post-booking automation. The least glamorous job and the one closest to revenue. Our workflows read call transcripts and trigger what happens next: recap, assets, CRM updates, follow-up sequencing. Most pipeline leaks between interested and signed, and this is the machinery that stops the leak.

Notice what is missing. None of these jobs is “build a 13-step GTM workflow.” The workflow is a means. Booked calls are the output. A GTM engineer who reports in workflows shipped instead of calls booked is a hobbyist with a salary.

The intent-signal stack, and what the tools are actually for

Ours: Clay for enrichment and waterfall lookups, Apify for scraping social activity, ad libraries for spend intelligence, hiring data for growth signals, Claude Code stitching it together. Swap any brand for its competitor and nothing important changes. The stack is commodity. The targeting logic is not.

The point of every tool in it is reaching people who are already moving. Of the 3,000+ calls we booked in the last 12 months, a vast majority came from prospects showing intent, not from people sitting on a scraped list.

What signal-based targeting looks like with real thresholds:

→ A LinkedIn lead is worth contacting at 5,000 followers minimum, 10,000 to be safe. Below that, “active on LinkedIn” is noise.

→ The hiring math: 500 to 1,000 companies hiring for a role your service replaces or supports converts to 25 to 50 qualified calls at a 5% rate.

That second play produced 24 booked calls in 30 days for one client without touching Apollo or ZoomInfo. Social activity, ad spend, and hiring signals did all of it.

One caveat before you build anything: intent targeting falls apart when your market is too small to hit volume. I’d rather you know that now.

Workflow theater: the part every guide skips

Tools without offer and list precision are workflow theater. The hype of Clay hyper-personalization is complete BS in my humble (yet correct) opinion. You can build a 40-node enrichment flow that writes a custom first line about the prospect’s dog, and it will lose to a plain email sent to a precisely defined list carrying an offer that sells to cold traffic.

The receipts: 587 positive replies from one ideal buyer profile in a single month, for a biz-op offer, with zero Clay. The list definition did the work (“young male founders, 20 to 35, in North America, going through the grind alone”) and the email stayed plain.

So run this test before any GTM engineering effort, in-house or hired. 2 things have to exist before automation earns its keep:

1. An offer a stranger would take.Not the offer that closes your referrals. That one assumes trust a stranger doesn’t have.
2. A list defined around a human, not a category.“B2B founders doing 30K a month” is not a defined list.

Automate before those exist and the system gets you to zero faster.

GTM engineer vs RevOps

The cleanest split holds up: RevOps runs the system of record, GTM engineering builds the system of action.

Honest addendum: at most companies under $25M in revenue this is one person or one partner wearing both hats, whatever the title says. The two-department version of the debate is a scale problem most readers of this post do not have yet.

GTM engineer salary: the real numbers

Straight from the job boards as of mid-2026:

→ ZipRecruiter puts the typical US range at roughly $140K (25th percentile) to $257K (75th percentile).

→ Glassdoor averages base pay near $131K.

→ Venture-backed B2B companies pay $130K to $260K base plus equity for senior hires, and the premium tracks technical depth. Builders who write code out-earn button-clickers by a wide margin.

Before posting the job, run the same math you’d run on an SDR. An SDR costs $85-150K a year, ramps for three to six months, works one channel, then reports that “the leads were bad.” A GTM engineer costs more and can produce far more, but only against enough market. Below roughly 100,000 reachable buyers there isn’t enough volume to find the winning offer, and no salary fixes that.

Fractional GTM engineer, GTM engineering agency, or full-time hire

3 ways to buy the capability, compared honestly:

A fractional GTM engineer. A senior builder on 10 to 20 hours a week. The right default for most companies, because the work is front-loaded: infrastructure, lists, and workflows get built once, then tuned. Paying a full-time salary through the tuning phase is how workflow theater gets funded.

A GTM engineering agency. The category runs from serious operators to workflow shops with a Clay certification. The tell is what they report on. Ask what happened to the last 20 calls they booked, not for screenshots of a workflow canvas.

A full-time hire. Justified once outbound is a proven channel and signal volume demands a dedicated builder. For most companies under $25M it is the third move, not the first.

Our own answer is a fourth structure: we build the whole machine in 90 days as a growth partner, run it, and hand over the keys, because buy vs build was always a false choice. Whichever route you pick, the sequencing rule doesn’t change: offer first, list second, automation third.

The guardrails nobody puts in the job description

A GTM engineer’s least visible job is keeping the sending machine deliverable. Google’s email sender guidelines require authentication (SPF, DKIM, DMARC), one-click unsubscribe, and spam complaint rates held under 0.3%, and Microsoft enforces its own version. Volume without deliverability infrastructure is self-sabotage. The spray-and-pray setups now land in spam and burn their own domains.

For scale reference: a mature campaign in our operation runs around 10,000 emails a day across a 60/40 split of Microsoft and Google inboxes, and the ramp to that number is engineered, not guessed. I broke down the sending math in how many cold emails to send per day, and the performance bar the whole system answers to in cold email reply rate benchmarks.

tl;dr: GTM engineering is real work wearing a faddish title. The job is removing human bottlenecks between a buying signal and a sent email. The tools are commodity, the salaries are public, and the whole thing lives or dies on two unfashionable inputs: an offer that sells to cold traffic and a list built around a human. Get those right and the engineering compounds. Get them wrong and you’ve automated a zero.

GTM engineering FAQ

More field notes like this live on the blog index.

And if you’d rather not referee the fractional vs full-time vs agency debate at all: this is the machine we build for B2B service businesses in 90 days, and you own it when it’s done. Book a calland we’ll tell you straight whether your market has enough signal volume to be worth engineering. Some markets don’t, and hearing that for free beats paying $180K to find out.